7654
Biochemistry 1987, 26, 7654-7667
Sutoh, K., & Lu, R. C. (1987) Biochemistry 26, 4511-4516. Sutoh, K., Yamamoto, K., & Wakabayashi, T. (1984) J . Mol. Biol. 178, 323-339. Sutoh, K., Yamamoto, K., & Wakabayashi, T. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 212-216. Sutoh, K., Tokunaga, M., & Wakabayashi, T. (1987) J. Mol. Biol. 195, 953-956. Szilagi, L., B a h t , M., Stretrer, F. A., & Gergely, J. (1979) Biochem. Biophys. Res. Commun. 87, 936-945. Tao, T., & Lamkin, M. (1981) Biochemistry 20, 5051-5055. Tokunaga, M., Sutoh, K., Toyoshima, C., & Wakabayashi, T. (1987) Nature (London) 329, 635-638. Tong, S. W., & Elzinga, M. (1983) J . Biol. Chem. 258, 13 100-1 3 1 10.
Towbin, H. M., Staehelin, T. M., & Gordon, J. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 4350-4354. Vanaman, T. C., & Stark, G. R. (1970) J . Biol. Chem. 245, 3565-3573. Walker, J. E., Saraste, M., Runswick, M. J., & Gay, N. J. (1982) EMBO J . 1 , 945-951. Warrick, H. M., De Lozanne, A., Leinwand, L. A., & Spudich, J. A. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9433-9437. Weeds, A. G., & Taylor, R. S. (1975) Nature (London) 257, 54-56. Yount, R. G., Okamoto, Y., Mahmood, R., & Grammer, J. (1985) Abstracts of the 1 lth Yamada Conference, Kobe, Japan, Aug 1985, p 41.
Reactions of the Sarcoplasmic Reticulum Calcium Adenosinetriphosphatase with Adenosine 5'-Triphosphate and Ca2+ That Are Not Satisfactorily Described by an El-E2 Modelt Neil Stahl and William P. Jencks* Graduate Department of Biochemistry, Brandeis University, Waltham, Massachusetts 02254 Received April 1, 1987; Revised Manuscript Received July 7, 1987
ABSTRACT: Phosphorylation of the sarcoplasmic reticulum calcium ATPase, E, is first order with
kb = 70
f 7 s-l after free enzyme was mixed with saturating A T P and 50 p M Ca2+;this is one-third the rate constant
of 220 s-l for phosphorylation of enzyme preincubated with calcium, CE.Ca2,after being mixed with ATP under the same conditions (pH 7.0, Ca2+-loadedvesicles, 100 m M KCl, 5 m M MgZ+,25 "C). Phosphorylation of E with A T P and Ca2+ in the presence of 0.25 m M A D P gives -50% E-P.Caz with kobsd= 77 s-l, not the sum of the forward and reverse rate constants, kobsd = kf k, = 140 s-l, that is expected for approach to equilibrium if phosphorylation were rate limiting. These results show that (1) kb represents a slow conformational change, rather than phosphoryl transfer, and (2) different pathways are followed for the phosphorylation of E and of C E C a 2 . The absence of a lag for phosphorylation of E with saturating A T P and Ca2+ indicates that all other steps, including the binding of Ca2+ ions and phosphoryl transfer, have rate constants of >500 s-l. Chase experiments with unlabeled ATP or with ethylene glycol bis(Saminoethy1 ether)-N,N,N',N'-tetraacetic acid (EGTA) show that the rate constants for dissociation of [ V - ~ ~ P ] A T and P Ca2+are comparable to kb. Dissociation of ATP occurs at 47 s-l from E.ATP.Ca2+ and at 24 sS1 from E-ATP. Approximately 20% phosphorylation occurs following an EGTA chase 4.5 ms after the addition of 300 p M A T P and 50 p M Ca2+ to enzyme. This shows that Ca2+ binds rapidly to the free enzyme, from outside the vesicle, before the conformational change ( k b ) . The fraction of Ca2+-freeE s [ ~ - ~ ~ P ] Athat T Pis trapped to give labeled phosphoenzyme after the addition of Ca2+ and a chase of unlabeled A T P is half-maximal at 6.8 p M Ca2+,with a Hill slope of n = 1.8. The calculated dissociation constant for Ca2+ from E-ATP-Ca, M2 is -2.2 X = 15 pM). The rate constant for the slow phase of the biphasic reaction of E-PCa, with 1.1 m M A D P increases 2.5-fold when [Ca2+]is decreased from 50 p M to 10 nM, with half-maximal increase a t 1.7 p M Ca2+. This shows that Ca2+ is dissociating from a different species, aE.ATP.Ca2, that is active for catalysis of phosphoryl transfer, has a high affinity for Ca2+, and dissociates Ca2+ with k I 45 s-l. It is concluded that steady-state turnover of the ATPase under most conditions occurs through the E.ATP.Ca2 pathway, which has a relatively low affinity for Ca2+, not the pathway through CE-Ca2(or ''E1-Ca2"). This results in 11-17% unphosphorylated enzyme in the steady state at saturating [ATP] and [Ca2+] because the kb step is partly rate limiting. The two pathways for phosphorylation can result in nonlinear Lineweaver-Burk plots for A T P and initial overshoots of phosphoenzyme levels.
+
R e a c t i o n s of the calcium ATPase (E)' of sarcoplasmic reticulum with its substrates ATP and Ca2+can occur by two pathways, depending on the concentrations of ATP and Ca2+. The upper pathway in eq 1 is the well-known pathway in which calcium binds first and causes a conformational change before
1 cEATP
'Ecaz
7
Cap
5'0wz E
EATP
hp aEATP
Y
,-" &E;
+
ADP
(1)
h- P
Ca2
'This is Publication No. 1614. This research was supported in part grants from the National Institutes of Health (GM20888) and the National Science Foundation (PCM-8117816). by
phosphorylation by ATP; the initial product ofthis reaction is designated CECa2simply to indicate that it is the stable form
0006-2960/87/0426-7654$01.50/00 1987 American Chemical Society
VOL.
of the enzyme with bound calcium. Binding of ATP to CE-Caz causes a second conformational change, with a rate constant of kd = 220 s-l, which converts the enzyme to a catalytically active form, "E.Ca,.ATP, that is phosphorylated very rapidly with a rate constant of k, 1 1000 s-l (Petithory & Jencks, 1986). The lower pathway in eq 1 is followed when ATP binds first or ATP and calcium are added together to the enzyme. We describe here the properties of this pathway, which are quite different from those of the upper pathway. It is almost certainly the pathway that is followed in vivo, where the ATP concentration is well above millimolar (Veech et al., 1979) and the calcium concentration changes from less than 0.1 pM to the micromolar range during muscle contraction (Tanford, 1981). It is well-known that ATP increases the rate of a conformational change associated with calcium binding (Sumida et al., 1978; Takisawa & Tonomura, 1978; Scofano et al., 1979; Inesi et al., 1980; Guillain et al., 1981; Pickart & Jencks, 1984; Fernandez-Belda et al., 1984). It has been shown that this activation involves ATP that is bound to the catalytic site and phosphorylates the enzyme with a rate constant of 70 s-I upon addition of calcium; there is evidence suggesting that this rate constant represents a conformational change (Stahl & Jencks, 1984). Similar behavior has been observed with bullfrog sarcoplasmic reticulum by Ogawa and Harafuji (1986), and a model for ATP activation has been proposed by these workers that has similarities, but also significant differences, compared with the model proposed by us. The results reported here provide further support for the assignment of this conformational change to the kb step in the lower pathway of eq 1. Models are essential for developing an understanding of complex systems such as the Ca and the Na,K-transporting ATPases, and the El-E2 model shown in Scheme I (and the similar E-E* and the E-E' models) has played an important role in the development of our present understanding of the mechanism of these enzymes (de Meis & Vianna, 1979; Siege1 & Albers, 1967; Post et al., 1969; Glynn & Karlish, 1975). However, models have a tendency to take on a life of their own and may be mistaken for reality; this can impede understanding. Furthermore, the properties of the intermediates in these models are not always clearly defined, and some of them have never been observed. The results reported here and other data in the literature are not easily described by the El-E2 model. We suggest that the properties of these enzymes should be described by a notation that is based simply on the chemical composition or a single, defined activity of the different enzyme forms. MATERIALS AND METHODS Materials. Reagents were generally of the highest purity available and were used without further purification. Na2ATP was obtained from Boehringer Mannheim (Sonderqualitat), and [ T - ~ ~ P I A T(>98% P purity) was purchased from New
' Abbreviations: SR, sarcoplasmic reticulum; SRV, sarcoplasmic reticulum vesicles; E, calcium adenosinetriphosphatase; EGTA, ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid: MOPS, 4morpholinepropanesulfonic acid; ATP*, [~I'~P]ATP; PEP, phosphoenolpyruvate: Tris, tris(hydroxymethy1)ammomethane.
26,
NO.
24,
1987
7655
England Nuclear. Tightly sealed sarcoplasmic reticulum vesicles were prepared from rabbit skeletal muscle by a slight modification of the MacLennan procedure, as described previously (Pickart & Jencks, 1982). The preparations hydrolyzed ATP at 3-5 pmol (mg of total protein)-' min-' under the standard conditions described below when the vesicles were made permeable to calcium ions with 4 pg/mL of the calcium ionophore A23 187. The total amount of phosphoenzyme, [E,,], that was observed for intact vesicles with saturating [Ca2+]and [ATP] was 2-3 nmol/mg total protein. Methods. All experiments were performed at pH 7.0, 0.1 M KCl, 5 mM MgS04, and 25 OC unless otherwise noted. Ca-ATPase activity was assayed spectrophotometrically by coupling ADP production to NADH oxidation with pyruvate kinase and lactate dehydrogenase (Rossi et al., 1979). Standard conditions were 40 mM MOPS, 100 mM KC1, 5 mM MgS04, 0.41 mM CaC12,0.40 mM EGTA (23 pM free Ca2+), 1.5 mM ATP, pH 7.0, and 25 "C. Protein concentrations were determined by the procedure of Lowry et al. (1951), with bovine serum albumin as standard. Concentrations of free calcium were calculated from a dissociation constant of Kdiss= 7.4 X lo-' M for CaeEGTA (Godt, 1974). This value was chosen because it was measured directly under conditions of pH, [KCl], and [Mg2+]identical with ours. It is nearly 4 times larger than the value of 2.0 X lo-' M derived from the constants of Schwarzenbach (Schwarzenbach et al., 1957) and is up to twofold larger than some values used in the literature (Guillain et al., 1980; Ogawa, 1968; Allen, 1977). Therefore, it is necessary to divide the [Ca2+]in this work by 4 or -2 for comparison with results obtained from the smaller values of &iss. The free Ca2+ concentrations were not corrected for the calculated changes resulting from changes in ATP concentration because these changes are smaller than the estimated error of the experimental data. The release of protons from EGTA upon formation of the Ca-EGTA complex was neutralized with 1.47 equiv of KOH added with the Ca2+when high concentrations of Ca2+ and EGTA were mixed. The formation or decay of phosphoenzyme was followed with a rapid-mixing apparatus that can be used with either three or four syringes of equal volume, as described previously (Stahl & Jencks, 1984). For three-syringe experiments, the temperature-equilibrated contents of syringes A and B are forced through a mixing block connected to a length of narrow-bore Teflon tubing for a time tl before being quenched in a second mixing block with hydrochloric acid from syringe C. For four-syringe experiments, another reactant pushed from syringe C is allowed to react for a time t 2 before being quenched with acid from syringe D in an additional mixing block. The quench solution contained 1.5 M HCl and 40 mM KH2P04for three-syringe experiments or 2 mM HCI and 55 mM KH2P04for four-syringeexperiments. The reaction times were calibrated from measurements of the hydrolysis of 2,4dinitrophenyl acetate by hydroxide ion (Barman & Gutfreund, 1964). Measurement of [32P]E-P.Ca2. Bovine serum albumin was added to the quenched reaction mixtures (0.30 mg/mL final concentration), followed by trichloroacetic acid at a final concentration of 12%, and the amount of [32P]E-P.Ca2 was determined essentially as described by Verjovski-Almeida et al. (1978). Passively Loaded SRV. Passively loaded SRV were used in some experiments in order to inhibit the hydrolysis of phosphoenzyme and permit accurate end-point determinations
7656
STAHL AND JENCKS
BIOCHEMISTRY
Scheme I1
Scheme 111 E*ATP*
hgCCa*+l , ' k-Q
'-ATP/
E-P
kt
",TP
=
kd
CkATP
kd + ' ~ - A T P k'ATPIATPl kb
+
'E-
P**Ca2
E*Ca2
The term NBis the net rate constant through the branched pathway for the disappearance of ECa,:
aE-P*Ca2
in the presence of a nonradioactive ATP chase. SRV were passively loaded with CaZ+by incubation for 4-16 h at 4 OC in a solution containing 13-60 mg/mL SRV, 0.1 M KCl, 5 mM MgS04, 5 mM MOPS, pH 7.0, 0.32 M sucrose, and 20 mM CaCl,, unless indicated otherwise. For each reaction, 10 pL of this stock solution was diluted into 0.99 mL of the reaction solution described for syringe A. This solution was loaded into syringe A of the rapid mixer, and the reaction was started within 15 s. This procedure results in the addition of 0.2 mM CaC1, in the A syringe. Leaky Vesicles. SRV were made permeable to Ca2+ by incubation at room temperature in the presence of 118 mM Tris (pH 9.0-9.5) and 1 mM EGTA. The length of time required to achieve leakiness varied with different enzyme preparations and the concentration of SRV. Vesicles were judged to be leaky when the addition of A23187 to the standard assay failed to cause a rate increase. The leaky SRV were neutralized to pH -7 with aqueous maleic acid. Typically, 1 mM Ca2+was added to help stabilize the leaky enzyme. Leaky SRV stored on ice showed no evidence of becoming resealed with time. Simulations. Simulations of the pre-steady-state kinetics were calculated with a program written for IBM-compatible microcomputers that allows input of rate or equilibrium expressions and the initial concentrations of each species for any model. The time course of a reaction is simulated by iterative calculation of the changes in the concentrations of each species over very small time increments, usually I 1 ps. The time increment used in the simulation of a model was 10.01k-', where k is the fastest rate constant in the model. The results can be shown on a graph or fit by a nonlinear least-squares procedure for estimation of rate constants. Rate Equation for Scheme II at Saturating [Ca2+].A rate equation used for the simulation of Scheme I1 at saturating [Ca2+]was derived by Cleland's method of net rate constants (Cleland, 1975). Scheme I1 describes the reaction sequence at saturating [Ca2+]because the second-order binding of Ca2+ is rapid. The net rate constant for a reversible step is calculated from the rate constant for formation of the intermediate multiplied by the fraction of the intermediate that partitions forward. This fraction is equal to the rate constant for that step divided by the sum of the rate constants for every pathway through which the intermediate can decay. The velocity at a given enzyme concentration, v / [E,,,], is equal to the reciprocal of the sums of the reciprocals for every net rate constant in the pathway. The velocity at each substrate concentration was calculated with a computer program from the expressions for the net rate constants shown below, the rate constants in Table 11, and estimated values of CkATP = lo7 M-I s-l and kLTP = 5.3 X IO6 M-' s-I. The net rate constants for reversible steps are designated by the variable N: 'NATP=
kb
1
k;Tp
E \
-
E*ATP**Ca2
kb
kLATP
The velocity at a given enzyme concentration was calculated from 1 = l/k, l/kt + l/kh + ~ / N B
+
Equation for Trapping E-ATP. The dependence on [Ca"] of the fraction of E.ATP* trapped in a pulse-chase experiment, in which unlabeled ATP and calcium are added to E.ATP*, may be calculated according to Scheme 111. A computer program that minimizes the sums of the squares of the residuals was used to obtain k3, k-3, and kLATPfrom the equations shown below. The other rate constants are taken from Table 11. The net rate constant for the reversible step in this scheme is k3[Ca2+]2(k!ATp kb) N3 = k-3 kb k!ATp
+
+
The fraction trapped Ft can be calculated from the product of the fractions partitioning toward formation of phosphoenzyme at each step: N3
Ft = N3
+ k-ATP
kb
kb
+ klATP
RESULTS Phosphorylation after Preincubation with EGTA. Figure 1 shows that the formation of [32P]E-PCa2 occurs with an observed rate constant of 62 s-' if passively loaded SRV are preincubated with EGTA for 15 s before mixing with 200 pM [ T - ~ ~ P I A Tand P 50 mM Ca2+ (dotted line). An equally good fit is obtained with two consecutive first-order rate constants of 67 s-l and 1060 s-l (solid line; see below). An experiment similar to that shown in Figure 1 but with leaky vesicles, 100 pM Ca2+, and 200 pM ATP gave the same initial rate of phosphorylation but a larger first-order rate constant of 84 s-l for approach to the steady-state concentration of phosphoenzyme (not shown). This rate constant is the sum of the first-order rate constants for the formation and disappearance of phosphenzyme (Hiromi, 1979). A similar rate constant of 77 s-I is observed for phosphorylation to reach equilibrium in the presence of 0.25 mM ADP and 0.5-1.0 mM [y-32P]ATPwith passively loaded SRV (Figure 2, solid lines). At these concentrations of ADP and ATP, the equilibrium amount of E-P-Ca, formed is 0.46-0.53 of that observed in the absence of ADP. Preincubation with the calcium ionophore A23 187 results in a 24-30% decrease in the steady-state level of [32P]E-P.Ca2 when passively loaded SRV react with 200 pM [T-~'P]ATP and 100 pM Ca2+for 2 s (Table I). The magnitude of the decrease is independent of the zmount of ionophore added over a fourfold range. This decrease in the steady-state concentration of [32P]E-P.Ca2 cannot be explained by the buildup of ADP because the addition of 9 pM ADP, which is an upper limit for the amount that can be formed in 2 s (calculated from
VOL. 26, NO. 24, 1987
SARCOPLASMIC RETICULUM CALCIUM ATPASE
,
I .O r
08
i V
I
1
I
I
IO
20
30
40
50
msec FIGURE 1: Formation of E-P.Ca2 by reaction of E with 200 pM [y-32P]ATPand 50 pM CaZ+. Final conditions were 40 mM MOPS (pH 7.0), 100 mM KCI, 5 mM MgS04, 2.5 mM EGTA, 2.41 mM CaClz(50 pM free calcium), 200 pM [y-32P]ATP,and 0.065 mg/mL SR protein at 25 'C. Syringe A contained 5.0 mM EGTA and 0.13 mg/mL SR protein; syringe B contained 4.82 mM CaCI, and 400 pM [y-'?]ATp; syringe C contained 1.5 N HCI and 40 mM KH2P04. Other components were present in all syringes except C at their final concentrations. SRV were passively loaded with CaZ+as described under Methods. A zero time point was measured by reversing the order of addition of syringes B and C. The best fit to a single exponential has a rate constant of 62 s-' (dotted line). The solid line is calculated for two consecutive exponentials with rate constants of 67 s-l and 1060 s-'.
7657
Table I: Amount of E-PCa2 Observed in the Presence and Absence of the Ionophore A23187" additions [E-P]/[E,,,] additions [E-PI / [E1011 none (1 .O) A23187 (49 mg/mL) 0.76 9 pM ADP 0.87 A23187 (25 mg/mL) 0.70 DMSO only 1.0 A23187 (12 mg/mL) 0.73 a Final conditions were 0.1 M KC1, 40 mM MOPS (pH 7.0), 5 mM MgSO,, 200 fiM [y-32P]ATP,and 100 pM Ca2+. One milliliter of a solution containing 0.3 mg of passively loaded SRV, 0.1 M KC1, 40 mM MOPS (pH 7.0), 5 mM MgSO,, 1 mM EGTA, and 1.1 mM CaC1, was incubated with the indicated additions for 15 s. Then, the reaction was started while vortexing by the addition of 1 mL of a solution containing 400 pM [Y-'~P]ATP,l mM EGTA, 1.1 mM CaCl,, 0.1 M KC1, 40 mM MOPS (pH 7.0), and 5 mM MgSO,, followed by the standard quench solution after 2 s. The volumes of the indicated additions were 5 p L of a 3.6 mM ADP solution and 2 pL for the DMSO or DMSO + A23187 solutions. The concentration indicated in the table is that for the stock A23187 solutions made in DMSO. The amount of E,,,, 2 nmol/mg, was taken from the amount of [32P]EPCa2 in the absence of any additions.
c 0 c
c
& 0.6 \
m
n I W
0.4
Y
0.2
20
12
24
36
48
60 96
120
msec FIGURE 2: Phosphorylation of E to give E-P.Ca2 wiih 0.5 (0),0.8 ( O ) , and 1.0 (A) mM ATP in the presence of 0.25 mM ADP. Conditions as in Figure 1 except syringe A contained 0.13, 0.27, or 0.27 mg/mL SR protein and syringe B contained 1.O, 1.6, or 2.0 mM [y-3ZP]ATPand 0.5 mM ADP. The solid lines are drawn for first-order reactions with a rate constant of 77 s-l and end points of O.46[qa] or 0.53[&], while the dotted line is drawn for a rate constant of 140 s-I. the measured ATPase activity at 1 mM ATP) accounts for only half of the observed decrease (Table I). We conclude that at least 11-17% of the enzyme is unphosphorylated during steady-state turnover in the presence of 100 pM Ca2+and 200 pM ATP. Trapping of E.ATP*. When passively loaded SRV were preincubated for 15 s with [ T - ~ ~ P I A Tand P EGTA, followed by the addition of 50 pM Ca2+ and a chase of nonradioactive ATP, about 55% of the E-P-Ca2 formed contains 32P(Figure 3, closed circles). The initial rates are the same, but the rate constant observed for phosphorylation in the presence of the nonradioactive ATP chase is 117 (closed circles), compared with 70 s-l in the absence of the ATP chase (open circles). These results show that the rates of phosphorylation and dissociation of Ca2+ from EsATP-Ca, are comparable.
40
60
EO
100
msec FIGURE 3: Formation of E-P*-Ca2 from E.ATP* with 50 pM Ca2+ in the presence of a nonradioactive ATP chase. Final conditions were 40 mM MOPS (pH 7.0), 100 mM KCI, 5 mM MgSO,, 2.5 mM EGTA, 2.41 mM CaCI2, 150 (0) or 30 pM ( 0 )ATP, and 0.065 mg/mL SR protein at 25 OC. Syringe A contained 5.0 mM EGTA, 300 (0)or 60 pM ( 0 )[y-32P]ATP,and 0.13 mg/mL passively loaded SR protein; syringe B contained 4.82 mM CaCI2and for the points marked ( 0 )2 mM nonradioactive ATP, 500 pM PEP, and 0.1 mg/mL pyruvate kinase. Syringe C contained 1.5 N HCI and 40 mM KH2P04. Other components were present in all syringes except C at their final concentrations. The regenerating system in syringe B is required to remove ADP contamination from the ATP used as a chase. The points marked (0)are from Figure 1 of Stahl and Jencks (1984). The line drawn through the closed circles is for a first-order reaction with a rate constant of 117 s-' and an end point of 0.56[Et,,]. The fraction of E.ATP* trapped decreases as the concentration of free Ca2+in the ATP chase is lowered (Figure 4). Passively filled SRV were mixed with 15 FM [ T - ~ ~ P ] A TinP the presence of 5 mM Mg2+and 2 mM EGTA for 100-120 ms to form E-ATP* and then mixed with 0.5 mM nonradioactive ATP and various concentrations of free Ca2+for 1 6 0 ms to allow phosphorylation to reach completion. Trapping of E.ATP* is half-maximal at 6.8 pM Ca2+,and a Hill plot of the fraction trapped against Ca2+concentration has a slope of 1.8 (Figure 4, inset). The line drawn through the points of Figure 4 was calculated from the equations describing trapping for Scheme I11 (Materials and Methods). The dissociation of [y-32P]ATPfrom E.ATP* occurs with a rate constant of 24 s-l in the presence of 5 mM Mg2+(Figure 5). This was measured with the four-syringemixing apparatus by preincubating passively loaded SRV with 15 pM [ T - ~ ~ P ] ATP in the presence of 5 mM EGTA and 5 mM Mg2+ for 15 s to form E.ATP*, then mixing with an excess of unlabeled ATP in the absence of Ca2+for various times t , , and finally
7658
BIOCHEMISTRY
STAHL AND JENCKS
n
I W u
20
40
60
80
100
tl,msec
0
1
I
I
I
1
3
6
9
12
15
[co2+1 , p~
4: Dependence on [Ca2+]of E.ATP* trapping to give EP*Ca2. With the four-syringe system, E was mixed with 15 pM [T-~~P]ATP for tl = 100-120 ms and then reacted with various concentrations of Ca2+in the presence of a chase of 0.5 mM nonradioactive ATP for t2 2 60 ms. All syringes except D contained 40 mM MOPS (pH 7.0), 0.1 M KCI, and 5 mM MgS04. In addition, syringe A contained 0.13 mg/mL SRV passively loaded with 50 mM CaCI2and 2.0 mM EGTA; syringe B contained 30 pM [Y-~~PIATP; syringe C contained 1.5 mM nonradioactive ATP, 19 mM EGTA, and either 0, 11.57, 14.83, 16.35, 17.23, 18.21, 18.75, 19.28, or 20.45 mM CaCI,. This rotocol was chosen in order to minimize the transiently high [Cap+]that would result from mixing together a high concentration of free Ca2+with EGTA since chelation is not instantaneous (Smith et al., 1977). For this mixing protocol, the initial Ca2+concentration calculated to exist immediately upon mixing during t2 for each point in the plot (the final [Ca2+]is shown in parentheses) is 0.37 pM (1.0 pM), 0.87 pM (2.0 pM), 1.5 pM (3.0 pM), 2.3 pM (4.0 pM), 5.6 pM (6.0 pM), 15.3 pM (8.0 pM), and 107 pM (12.0 pM). The amount of [32P]E-P~Ca2trapped at each [Ca2+]was calculated as the fraction of [E-PI,,,, the amount trapped in a run with 300 pM Ca2+(not shown), which is plotted as 0.6[Et0,];the relative fractions trapped at lower [Ca2+]are adjusted accordingly. The solid line was calculated according to Scheme I11 (Materials and Methods). A Hill plot of the [Ca2+]dependence for the fraction of the maximum trapping of EeATP is shown in the inset. FIGURE
reacting with saturating Ca2+ for sufficient time to allow phosphorylation to proceed to completion. Trapping of E.ATP-Ca2. Figure 6 shows the observed formation of phosphoenzyme when passively loaded SRV are preincubated with EGTA for 15 s and then reacted with 300 pM [ T - ~ ~ P ] A Tand P 50 pM free Ca2+for tl = 4.5 ms, followed by a chase of 5 m M EGTA during t 2 that lowers the concentration of free Ca2+to 0.25 pM. The second-order rate constant of 2 X lo6 M-' s-l for chelation of Ca2+by EGTA (Smith et al., 1977) corresponds to a half-time of 0.1 ms for chelation by 5 m M EGTA. About 20% of the enzyme becomes labeled during t l , and 20% of the remaining enzyme becomes labeled following the EGTA chase. A similar experiment in which the ATP was present during the 15-s preincubation to form EoATP resulted in phosphorylation of 18% of the enzyme following the EGTA chase (data not shown). This shows that preincubation with ATP has no effect. The fact that the level of phosphoenzyme does not increase after 10 ms shows that the rate constant for phos-
FIGURE 5 : Measurement of the rate constant for the dissociation of E.ATP* (E.[r-"P]ATP). The rapid mix apparatus was used with four syringes. All syringes except D contained 40 mM MOPS (pH 7.0), 0.1 M KCI, and 5 mM MgSO4 In addition, syringeA contained 5 mM EGTA, 15 pM [T-~~P]ATP, and 0.13 mg/mL passively loaded SRV; syringe B contained 1 mM nonradioactive ATP; syringe C contained 5.9 mM CaCI2. This provides a concentration of free Ca2+ of 300 pM during t 2 . For the points at t l = 0, 5.6 mM CaCI, was added to syringe B, and syringe C contained only 0.3 mM CaCI,. The plot shows the fraction of [32P]E-P.Ca2 (compared with t l = 0) that remains after various times t,. The length of t2 was 250 ms to allow phosphorylation (which has an observed rate constant of 117 SKIin the presence of a nonradioactive chase, Figure 3) to reach completion. The line is drawn for a first-order rate constant of 24 s-l.
0.8
\
0.6 I
I
Q
0.2
12
24
36
48
60
t 2 , msec
FIGURE 6: Phosphorylation of E-ATP-Ca, to give E-PCa, in the presence of an EGTA chase. With the four-syringe system, E was mixed with 50 pM Ca2+and 300 pM [y-32P]ATPfor t l = 4.5 ms and then chased with 5 mM EGTA for various times 1, to lower the concentrationof free Ca2+to 0.25 pM. All syringes except D contained 40 mM MOPS (pH 7.0), 0.1 M KCI, and 5 mM MgS04. In addition, syringe A contained 0.61 mg/mL passively loaded SRV and 5 mM EGTA; syringe B contained 5.03 mM CaCI2and 600 pM [y-32P]ATP syringe C contained 15 mM EGTA. t l was held constant at 4.5 ms, and the duration of t 2 was varied. The fraction of E,,, that is phosphorylated at t2 = 0 was measured by reversing the order of addition from syringes C and D. The line is drawn for a first-order rate constant of 350 s-l as described in the text, starting at 0.20 and with an end point of 0.20 + 0.17 = 0.37.
phorylation of E.ATP* in the presence of 0.25 pM Ca2+ is insignificant. Dependence on [Ca2'] of the Reaction of E-P-Ca2 with ADP. The reaction of [32P]E-PCa2 with 1.1 m M ADP in the presence of a nonradioactive ATP chase is biphasic, proceeding with a rapid burst of [32P]E-PCa2 disappearance followed by a slow phase (Figure 7; Sumida et al., 1980; Froehlich et al., 1980; Pickart & Jencks, 1982; Froehlich & Heller, 1985; Wang, 1986; Fernandez-Belda & Inesi, 1986). The size of the burst is independent of the concentration of Ca2+, but the observed rate constant for phosphoenzyme disappearance in the slow phase increases 2.5-fold when the
SARCOPLASMIC RETICULUM CALCIUM ATPASE
VOL. 26, NO. 24, 1987
7659
I ,"'e Scheme 11'
0.0
1
2 ~ 0 " t 'Eca2
0.6
IO
20
30
50
40
msec
7: Effect of [Ca2+]on the reaction of [32P]E-P.Ca2 with 1.1 mM ADP. [32P]E-P.Ca2 was made by preincubating 0.29 mg/mL passively filled SRV with 10 pM [Y-~~P]ATP in 200 pM Ca2+ during the 15 s required to load the solution into the mixer. The reactions were carried out in the presence of 50 ( O ) , 5 (U), 3 (A), 1 ( O ) , 0.6 (X), 0.1 (v),and 0.01 (0) pM free Ca2+. In addition to 40 mM MOPS (pH 7.0), 0.1 M KC1, and 5 mM MgS04, syringe and passively filled SRV, and syringe A contained 10 pM [Y-~~P]ATP B contained 15 mM EGTA and 14.8 ( O ) , 13.0 (a), 11.9 (A),8.5 (0), 6.6 (X), 1.7 (v),or 0 ( 0 )mM CaCI,. FIGURE
l5l
DISCUSSION
H -- 0I 42
-040
040
I20
0
15
30
45
60
order to set limits for the specificities that are defined by the coupling rules for this enzyme. The experiments were carried out under standard conditions, except as indicated. An upper limit of kHOHI 0.14 s-l was obtained for the hydrolysis of E-P.Ca2 from an observed half-time of 5 s for the disappearance of labeled phosphoenzyme that was obtained by manual quenching. Intact vesicles that had been passively loaded with 20 m M Ca2+ were phosphorylated with 15 pM [y-32P]ATPand 300 p M Ca2+ and then added to 0.5 mM unlabeled ATP (which had been incubated with phosphoenolpyruvate and pyruvate kinase to remove contaminating ADP). An upper limit of 500 s-l (Figure 3; Stahl & Jencks, 1984). (7) The small second-order rate constant of 5 X lo3 M-' s-I for binding of Pi or Mg2+ to E (pH 6.0, [KCl] = 0, and 23 "C) is consistent with the suggestion that this binding involves a conformational change (Guillain et al., 1984). The formation of E-P-Mg from E involves a change in vectorial specificity that permits productive binding of Ca2+ to the phosphoenzyme from inside the vesicle. This change in the exposure of the calcium binding sites from outside to inside must almost certainly involve a conformational change.
u
L
outside
ri
H20 vectorial specificity c a 2 + on-off
7665
E-P
/ inside
4
2+
PCai,
\
ATP-ADP chemical specificity
Pi-H20
(8) The phosphorylation of aE-ATP.Ca2to give E-PCa, changes the direction from which calcium can dissociate, from the outside to the inside of the vesicle. This is the critical vectorial step for the active transport of calcium, and it must almost certainly involve a conformational change. Surprisingly, it occurs with a rate constant of >lo00 s-l, which is too fast to measure with presently available techniques (Stahl & Jencks, 1984; Petithory & Jencks, 1986). Coupling. The coupling of vectorial Ca2+transport to ATP hydrolysis results from changes in the catalytic and vectorial specificities of different enzyme states. A simple set of specificity rules that describe coupling may be taken from experimentally observed properties of the calcium ATPase (Jencks, 1983). These rules represent changes in chemical specificity for catalysis and in vectorial specificity for calcium binding in different states of the enzyme, as outlined in Scheme VIII. (1) The enzyme catalyzes reversible phosphorylation by ATP, not Pi, when Ca2+is bound to the transport sites. (2) The enzyme catalyzes reversible phosphorylation by Pi, not ATP, when the transport sites are not occupied by Ca2+. (3) Productive binding and release of Ca2+ occur only on the cytoplasmic side of the membrane when the enzyme is unphosphorylated. (4) Productive binding and release of Ca2+ occur only on the inside of the membrane when the enzyme is phosphorylated. Thus, Ca2+ binding acts as a catalytic switch that changes the catalytic specificity of the enzyme from Pi to ATP by facilitating an enzyme conformational change following the binding of ATP, while phosphorylation acts as a vectorial switch that changes the orientation of Ca2+binding and dissociation from cytoplasmic (outside) to inside the vesicle. Violation of any of these rules results in uncoupling of Ca2+transport from ATP hydrolysis; there is no one step in which coupling can be said to occur. The rules describe a strictly ordered kinetic mechanism with alternating chemical and vectorial steps. The cleavage of ATP and the transport of calcium are each divided into two parts that are sandwiched between the parts of the other process, so that neither the chemical nor the vectorial reaction can occur without the other. Our results provide further support for the hypothesis that phosphorylation is the switch that determines on which side of the membrane Ca2+binds and dissociates. We have concluded that the net rate constant for e a 2 +dissociation from aE-ATP-Ca2has a value of 45 s-', According to Scheme VII, there are three possible pathways through which Ca2+ can dissociate from this intermediate: (1) direct dissociation from aE.ATP-Ca2 through the ak-3 step, (2) reversal of the kd conformational change to give CE.ATP.Ca2,and (3) reversal of the kb conformational change, to give E.ATPCa2. The observation that the increase in the net rate constant for Ca2+ dissociation is half-maximal at 1.7 pM Ca2+(Figure 4) shows that little if any of the dissociation occurs through (3) because trapping of E.ATP.Ca2 is half-maximal at 6.8 pM Ca2+. The Ca2+dissociation cannot be accounted for by ( 2 ) because the
7666
BIOCHEMISTRY
ratio of the rate constants for dissociation of Ca2+and ATP is 45/37 = 1.2 from "E-ATP.Ca2 and 80/120 = 0.67 from CE-ATP.Ca2(this work; Pickart & Jencks, 1982; Petithory & Jencks, 1986). Furthermore, dissociation of Ca2+ from CE. ATP-Ca, would require a rate constant of k,, = 310 s-l along with the measured value for Ck-3= 80 s-l (Petithory & Jencks, 1986) in order to account for a net rate constant of 45 s-l for CaZ+ dissociation through pathway (2). The reverse rate constant k,, is smaller than the forward rate constant kd = 220 s-' because CE.ATP-Ca,is observed to form aE.ATP.Ca2, with a rate constant of kd = 220 s-l in the absence of ADP and 270 s-' in the presence of ADP (Petithory & Jencks, 1986). We conclude that some or all of the Ca2+must dissociate directly from aE.ATP.Ca2. Even if all of the ATP dissociation were to occur through CE.ATP.Ca2,this would account for only 80/ 120 X 37 s-l = 25 s-l for Ca2+dissociation, so the Ca2+ must dissociate directly from aE.ATP.Ca, with ak-32 2 0 s-l. Thus, the phosphorylation step itself, with k, 1 1000 S K I , appears to be the critical vectorial step after which Ca2+ dissociation to the outside is forbidden and dissociation to the inside becomes possible. Coupling of the transport of two Ca2+ions to the hydrolysis of one molecule of ATP is remarkably strong; the Ca2+concentration gradient that accumulates at different ratios of ATP and ADP is experimentally indistinguishable from that predicted from the Gibbs energy of ATP hydrolysis (Trevorrow & Haynes, 1984). Our kinetic analysis of the extent to which an enzyme intermediate reacts in a manner that violates a coupling rule agrees with this conclusion. For example, the direct hydrolysis of E-PCa, violates the rule which states that the reaction of phosphoenzyme with water may occur only when Ca2+is not bound; it results in uncoupled hydrolysis of ATP, The upper limit of 10.14 s-l (Results) for this hydrolysis gives uncoupling of lo00 s-l for the reaction of E-PCa, with ADP (Petithory & Jencks, 1986). Phosphorylation of the enzyme by ATP in the absence of calcium to give E-P violates the other specificity rule and would also give uncoupled hydrolysis of ATP, because E-P reacts rapidly with water. The upper limit of
